More on uninsured bank deposits

One popular narrative about Silicon Valley Bank is that it was felled by nothing more than boneheaded management.

Now, the mismanagement part may be true, but it isn’t the entire story. A new paper from a trio of academics lays out an under-recognised element of the run on SVB, and the ensuing mini-crisis in the regional banks: The banks that failed, particularly Silicon Valley Bank and First Republic, weren’t holding on to uninsured deposits in concentrated sectors because they were dumb. It was part of their business strategy. And until recently they were rewarded for doing it.

As the professors — Briana Chang of the University of Wisconsin, Ing-Haw Cheng of University of Toronto and Harrison Hong of Columbia University and NBER — put it:

We propose a fundamentals-based role of uninsured depositors for regional banks. This view holds that SVB was at the tip of a systematic pattern of banks requiring large, uninsured deposits to support banking relationships underlying their specialised risk-taking strategies.

The specialisations of SVB and FRC (in tech start-ups and high-net-worth clients, respectively) were well known. It was also obvious that, as the professors establish, banks with the highest share of uninsured deposits faced greater trouble in the panic earlier this year.

But the consensus around management stupidity seems a bit . . . convenient, doesn’t it? Normally when a highly regulated, utility-type business fails, the explanation isn’t “listen, those guys were just real dimwits”. Even the Federal Reserve seems to have adopted an explanation that SVB was run by morons, saying that “senior leadership failed to manage basic interest rate and liquidity risk”.

In their paper, the academics find evidence that adds nuance to that view:

. . . during this pre-crisis period, banks with greater uninsured deposits were more profitable, valuable, and experienced inflows (Fact 3) despite their greater risk, suggesting that they have valuable underlying business strategies. Fact 4 indicates that uninsured deposits capture a dimension of risk beyond those captured by what is on the balance sheet. Fact 5 further supports the idea that risk-taking was part of a business strategy since such strategies at financial firms require high executive pay and strong incentives to execute in equilibrium.

Lest we forget, uninsured deposits are nearly free funding! Assuming a bank can hold on to them, at least.

The academics also argue that a decent share of SVB’s uninsured deposits were companies’ working capital, which is interesting because those deposits are (in theory) less likely to run.

We also have learned since SVB’s failure that the bank was, in fact, using a service to farm out deposits for FDIC insurance. And according to this Bloomberg story (with an accidental leak from the FDIC), one of the bank’s biggest depositor was . . . IntraFi, the company that manages deposits. Perhaps a VC-fuelled panic did in fact drive the sudden run on deposits by tech companies; companies otherwise have little reason to withdraw their working-capital funding, which they use for payroll, vendors and the like.

Also notable: Before 2022, the academics found “there is little evidence that greater uninsured deposits were associated with greater traditional balance sheet measures of risk.”

This finding raises more questions about whether it’s reasonable to rely solely on bank capital to offset financial-sector risk. Our colleagues at Unhedged have unpacked these questions nicely in recent days. These types of bank strategies don’t surface overnight, and it’s well known that SVB pursued a strategy of tech specialisation from the very beginning.

Again, the authors of the paper find evidence that these specialised banks (and executives, and shareholders) reaped rewards from their strategies before 2022, albeit with higher return volatility:

What’s more, the riskier banks’ executives were better paid, with boards that were more likely to be independent:

In a sense, this is just conventional wisdom. As the big global banks learned during the GFC, risk management isn’t good for business until it’s really necessary.

More interesting is the framework that the academics propose for what exactly changed between 2021 and 2022 that killed these banks’ business models: Interest rates. And not because rising interest rates create losses in bonds and loans, though that obviously contributed.

Instead, they propose a model where uninsured deposits are a key part of these banks’ businesses. Banks want cheap funding, as a general rule, and smaller banks have to pick specific spots to seek it out.

Specialisation should give banks better information about their chosen industry or region, and allow them to provide cheaper credit to risky borrowers. Those risky borrowers are then more willing to hold uninsured deposits at those banks. (It also seems reasonable to expect riskier borrowers to be more willing to stash huge amounts of uninsured cash with a bank.)

The problems arise when interest rates climb. That dries up funding for risky projects, and a broad range of depositors and investors can find higher-yielding safe options for their cash.

As the academics put it:

. . . the fundamentals-based view emphasises that the concentration of uninsured deposits at risk-taking banks such as SVB, First Republic, and others was not accidental . . . furthermore, that the fundamental origins of the crisis lay in a decline in match values due to rising interest rates and declining risky opportunities. While allowing for a role for contagion and bank runs to exacerbate these outflows, this view emphasises the equilibrium relationship between uninsured deposits and risk-taking. In plainer words, SVB was less an outlier, and more a fundamentally risky business whose uninsured depositors were a critical part of its franchise value.

In this context, one could argue that uninsured depositors benefited from financing risky banks — while they were alive, at least — without any consequences from their failure.

SVB and FRC shareholders, on the other hand, got an important lesson in a different risk-related metric: FAFO.